“…After screening the titles, 395 abstracts and 119 complete articles were reviewed, and 88 were selected for detailed evaluation. Of these, 59 were rejected for the reasons as follows: the study population was not a general population or the IRs were reported for NSAID users and non‐users, but no data were provided for estimating incidence among the general population ( n = 15); the outcome was UGIB, but the percentage of bleeding cases caused specifically by peptic ulcer was not provided ( n = 5); the same population was used by different studies ( n = 11), and only the most recent one is included; no IR estimates, person–time, or number of cases were available ( n = 13); the measure of frequency was prevalence rather than incidence ( n = 4); only mortality or operation rate was reported ( n = 9); and the outcome was DU or GU alone ( n = 1) . We excluded another study because it used a rough approximation of population at risk (denominator); moreover, it was the only eligible article from a non‐Western country .…”